Two-stage unscented Kalman filter algorithm for fault estimation in spacecraft attitude control system

Abstract

The study of fault/bias estimation based on the two-stage Kalman filter and the unscented Kalman filter in the presence of unknown random biases is addressed. Two kinds of faults are taken into account: additive faults and multiplicative faults, which are modelled as actuator faults and sensor faults in the spacecraft attitude control system (ACS). In accordance with the characteristic of the fault model of ACS, where the system state and the faults are decoupled, a novel two-stage unscented Kalman filter (TSUKF) algorithm is developed to estimate the decoupled states and biases simultaneously. By employing the unscented transform, the TSUKF algorithm does not need any linearisation of non-linear system models or the augmentation of the state, contributing to a more precise estimation. Meanwhile the computational cost is reduced by exploiting the bias-separate principle. The simulation results demonstrate the proposed algorithm when a micro-spacecraft is tracking a stable/manoeuvring target.